Presentation to American Paralysis Association Meeting May 20th, 1984 San Francisco, CA
Ney York University Medical Center, New York, NY
I would like to thank several individuals and the American Paralysis Association (APA) which made this symposium possible. I think this symposium is quite unique in this field and time, bringing together a collection of researchers in the area of CNS trauma research. I would like to particularly thank Mr. David Camhi, as a member of the Board of Directors of the APA who contributed a great deal to the laboratory at New York University Medical Center, encouraging us, giving us his hope and enthusiasm when ours have flagged in the course of the research we have done. I would also like to thank Admiral Van Orden for his role in making this symposium possible.
Finally, I would like to thank Dr. Joseph Ransohoff, the Chairman of my department, who first steered me into the area of spinal cord research. I was a lowly resident in Neurosurgery in his department when he encouraged me to do spinal cord injury research. I knew nothing about spinal cord injury, this was about seven years ago.
What I will describe to you today is the fruition of about 7 years worth of work in this field. I must also point out the work I do is not by any means done alone, there are three individuals in the group who have participated and contributed to all the work I will show you today. These are, Dr. Andrew Alight, who is a morphologist and a cellular neurophyiologist, Dr. John Gruner, who is responsible for a great deal of the animal behavior testing and Dr. Eugene Flamm, who is a neurosurgeon but has been very supportive and a source of many of the ideas we have looked at.
Last but not least, I would like to thank the Diapulse Corporation of America for having loaned us their machine, which sat in my laboratory for many years. I promised Mr. Bern Siler of the Diapulse Corporation that I would use the machine as soon as I saw some situation that warranted it. He waited patiently for three years before we started.
When I first entered this field of spinal cord injury, I was rather dismayed by the fact that everybody treated spinal cord injury as if it were a black box. People put treatments on as if it was "holy water" and they expected something to result from the treatment. Very often things did happen when you treated the spinal cord but very seldom did anybody have an explanation as to why recovery occurred or did not occur. Consequently. When I first started, I thought it was time to probe into the spinal cord, look at the phenomenology in the spinal cord and for a number of years we focused on the events in the first three or four hours after a contusion injury. We started out at this point because it seemed logical to do so.
Very little was in fact known for sure about the changes that occur, however, it's been known, and you heard it described on the first day of this meeting, when you hit a spinal cord sufficiently hard the animal becomes paralyzed within one or two minutes after you hit the cord. You look at the cord, histologically there appears to be very little damage. Then over the ensuing two to three hours, central hemorrhagic necrosis develops and the animal doesn't recover from it. The approach we took for a number of years was that we can perhaps stop the progression of event by introducing treatment shortly after the injury.
I wish to point out that this line of research, thought it may seem very much on the preventative side of spinal cord injury research, really is quite relevant to the work you have heard described yesterday and today. Virtually every one of the treatments we are now contemplating for spinal cord injury, which involves implantation of embryos, implantation of tissue and application of growth factors are invasive. That these methods produce further damage to the cord and some method must be available to prevent this damage or retard this damage before the treatments can be applied.
Slide 1:
This is a picture of a spinal cord and I show it just to remind you of some of the talks that have gone on in previous days. What you get after you hit a spinal cord, four hours after injury, is a loss of the central matter in the spinal cord. This is a longitudinal display and there is typically a preservation of a strip or rim of white matter on the outside rim of the cord. Right after you hit the spinal cord, these are somatosensory evoked potentials in our particular spinal injury model.
Slide 2:
Prior to injury you get a normal evoked potential. Right after impact, it's all flattened out (I don't show you the flattened response). At one hour we see a little recovery of the evoked potential. At three hours it is absent and at 24 hours it is gone.
Slide 3:
We were very interested in the course of events in the spinal cord during those first few hours that was responsible for this loss of action potential conduction across the lesion site. The approach we took was primarily to record and look at the blood flow in the spinal cord and at the same time we monitored blood flow with hydrogen clearance, we also measured extracellular ions. We measured extracellular calcium, potassium, and PH as we wanted to define the extracellular environment of the spinal cord during those first few hours.
Slide 4:
These were some of the first experiments that were performed three or four years ago, that were published. Essentially what happened, blood flow which started out at about 12 milliliters per hundred grams per minute, falls significantly by one or two hours after injury. Blood pressure during this time is recovering towards normal. We were very surprised to find, immediately after you hit the spinal cord, extracellular potassium which as you know is approximately 4 millimoles, rises very dramatically to a mean of 54 millimoles and then gradually clears out of the spinal cord over a period of three to four hours. Incidentally, there is a large standard deviation here because we believe the value of potassium reached actually goes close to 90 millimoles, but we impaled the spinal cord two to three minutes after the impact. If we get it into the spinal cord early enough, we will see the extracellular concentrations up in that region. What we found was the blood flow in the spinal cord as well as somatosensory evoked potentials (SEP) correlated very closely to the time course of extracellular potassium changes. In other words, just briefly describing the changes, the SEP's are gone with the high extracellular potassium and when extracellular potassium reaches somewhere in the order of 10 to 15 millimoles we see SEP's recovering in the animals. Blood flow did not fall significantly until extracellular potassium is returned to below 15 millimoles. Of interest is the fact that even though extracellular potassium does not rise again, SEP's disappeared and blood flow continued to fall.
Slide 5:
We then focused our attention on extracellular calcium and as you know, calcium is very rigorously excluded from the intracellular space by neurons and massive entry of calcium into neurons can kill them. What we found, when we recorded extracellular calcium which pre-injury is normally at 1.2 millimoles, was at the center of the contusion site extracellular calcium falls from 1.2 millimoles down to less than .01 millimole. In fact, it went below the calibrated range of our electrodes. We now estimate it to be in the region of 1 micromole.
In the area immediately surrounding the center of the contusion site, in the white matter, we find that calcium falls to very low levels, recovers, and then falls again. In the adjacent areas, calcium begins to recover slowly back towards normal. These findings strongly suggest that contusion injury causes entry of calcium into cells.
Slide 6:
We have also conducted similar experiments, looking at a variety of treatments, seeing how they effect the blood flow changes and the calcium changes...I won't go through all of them. What I want to do today is describe a fairly novel approach or treatment method we believe has not been tried before in spinal cord injury to look at these types of changes.
Slide 7:
Just briefly to orient you, the intracellular concentrations of potassium is on the order of 90 millimoles, sodium is about 37, free calcium inside cells is on the order of .001 millimole and it interacts with bound calcium in the tissue. Outside the cell, however, potassium is only 4 millimoles, sodium is 130 millimoles and calcium is 1.2 millimoles.
I've drawn arrows indicating the direction these ions will go if you disrupt the membranes. Indeed, our extracellular ionic studies suggest that when you hit the spinal cord there probably is as much as 80 to 90 percent disruption of cells at the impact site. The ions simply flow across membrane down to concentration gradients. We were very anxious to look at what the extracellular ionic changes were. As you know, there is no good way of distinguishing intracellular ions simply because you can't dissect out a single cell and look at what's inside it. So we chose the poor man's method of approaching this by simply measuring the total tissue ions.
Slide 8:
We rationalized that if we knew the extracellular ions and we knew the total tissue ions, by subtracting the two we could obtain some kind of estimate of the intracellular ionic levels. Let me just briefly run through this. What we do is take the spinal cord, cut it into two to three millimeter slices (we can cut it to much finer pieces), this is the lesion center, each piece is then dried so we can get dry/wet weights for water content then each piece is individually analyzed for all three ions; calcium, potassium, and sodium.
The next three slides you'll see represent the ionic changes along the length of the spinal cord in this manner, with the rostral cord in this direction and the caudal cord in that direction. It is a very straight forward analysis.
Slide 9A:
At one hour, what we find in the contused cord is a large rise in tissue sodium, a fall in tissue potassium, and an increase in tissue calcium. The changes are quite large (you probably can't see this), but potassium falls somewhere in the region of 80 millimoles to 40 millimoles. Fifty percent of the potassium is lost from the tissue altogether. This really tells you how massive the disruption is and the histological picture you see is very misleading. We believe now, close to 80 to 90 percent of the cellular volume of tissue is simply disrupted and the ions pour out.
Slide 9B:
When you go to 3 hours after injury, this is the picture you see and sodium and potassium have become substantially worse. What you find is calcium accumulating in contrast to the sodium and potassium changes, calcium is accumulating in the surrounding cord and you get two very large peaks of calcium. All these are standard deviation bars, all the numbers are quite significantly elevated from normal uninjured cords. At this time we begin to look for ways to manipulate the tissue calcium.
As you may be aware and certainly Dr. Bassett will summarize some of this information tomorrow for you, electromagnetic fields, particularly modulated electromagnetic fields have significant effects on tissue calcium. Ross Adey and many of his collaborators and several other laboratories have shown that modulated electromagnetic fields change calcium efflux from tissues.
Slide 10:
We have the Diapulse machine in our laboratory which puts out a paradigm of electromagnetic fields which is quite interesting. What the machine does is put out a very large field gradient of 47 volts per centimeter which is measured at the treatment head of the machine. It can be applied externally. The field has a carrier frequency of 27.12 megahertz, but if you apply that kind of electromagnetic field to tissue you would tend to heat it up. For this reason the field is pulsed at 65-microsecond pulse duration, 400 times per second. At this pulsation frequency, at these powers, this does not significantly heat tissue at all.
Slide 9C:
This is what we found the Diapulse machine did to our tissue; it had no effect on sodium, no effect on potassium, but it did significantly reduce those two large humps of tissue calcium immediately surrounding the impact site. We were intrigued by these results. The effect appears to be very specific for calcium because we see no changes in the sodium or potassium, furthermore, we see no changes in the water content of the spinal cord. Although many people speak of the edema in the spinal cord, there has been very little data presented changes early in spinal cord. I want to point out to you that in a contused cord, three hours after injury there is nearly a 6% increase in water content in the cord. This is an untreated animal versus pulsed electromagnetic field and we have done many experiments with other treatments using this paradigm.
Our question then became what effect does this pulsed electromagnetic field have on the recovery of the animals, if it does have this very specific effect on calcium. We realized at this time we were on to something that would not be readily accepted unless we did a "tour de force" of showing that the treatment had an effect. There was quite a bit of controversy about treatment, how to document it, so we have conducted quite an extensive series of experiments which are summarized here.
Slide 11:
These are the groups of animals we tested. Initially we did a series of ten animals treated with pulsed electromagnetic fields starting at one hour after injury and going for four hours. Then the animal was not treated any further. This is compared with 8 control animals.
We also compared four groups of animals where the treatment was limited to only one hour but started at 1 hour, 4 hours, and 24 hours after injury. These three groups of animals were treated daily for two weeks afterwards. As this progresses you will see the rational for this. We also decided to evaluate the animals in four different ways to make absolutely certain that the effects we are seeing would be real.
Slide 12:
First we use SEP's to monitor the animals, the paradigm is shown. The posterior tibialis nerve in these cats, this particular somatosensory ascending system goes through the dorsal column, going to the dorsal column nuclei then onto thalamus and is recorded from the contralateral cortex, where epidural screw electrodes are implanted permanently in the cats. We also monitor the median nerve evoked potentials in these animals as a control. The impact site is T7.
Slide 13
Briefly, the way we evaluate the evoked potentials, these are evoked potentials from one of the sets of control animals at 4 weeks after injury. As you see, it's virtually all noise. All these are evaluated double blind, we do not know which animals were treated until after all the data has been collated. I believe one of our observers felt that this evoked potential is significant but we saw nothing else here that was significant.
Slide 14:
Here is an example of some return of evoked potential and you can see the return in some of these animals but these are just to illustrate what we do.
Slide 15:
We were also, at the time, very concerned with monitoring descending function as well as ascending function. We took advantage of a very special reflex that cats and virtually all mammals have; that is, when you take an animal and you drop the animal into free-fall, the animal will exhibit a very stereotyped movement of all four limbs and hind limbs. This is a cat that has been sedated with ketamine. He has electrodes implanted into 12 muscles in the back of his legs plus we have electrodes implanted into the forelimbs, the biceps and triceps. When the animal is dropped, the data is collected on a computer. This is the kind of data that we can get on these cats. This is a normal cat, these are EMG's that we obtained from the cats. You can see there is a very fast response at about 18 to 20 milliseconds in the quadraceps of the left vastus, a response in the left gastroc, the right vastus, the right gastroc, right biceps and the right triceps.
Slide 16:
Immediately after injury what you find is a loss of the response in the lower limbs. This is a very reproducible response, the animals do not have to be treated or trained; we don't have to do anything. These kinds of data, we felt was easy to interpret and of course, the forelimb muscle remains intact.
Slide 17:
I want to go through some exambles of the changes we see and how it correlates with function just to convince you this test is quite reliable. Here I show it in a slightly different format. Here we show three muscle groups from each leg; all the left legs are here, all the right legs are here, the left vastus lateralus, the left anterior tibialus, the left gastroc and so forth, for the right leg. This is an animal that cannot walk after injury. This is a totally paralyzed animal. As you can see the animal has virtually no responses.
Slides 18, 20, 19:
This is an animal that cannot walk either, but this animal has some delayed and some bizarre responses occasionally.
This is an animal that we would regard to be a poor walker and that is an animal that can ambulate for fairly long distances but looks very abnormal during the walking. You can see the presence of some responses in the quadraceps, on both sides occasional responses in the gastrocs and tibialus and so forth. This is an animal that has been treated, this animal is a good walker and could be barely distinguished from normal.
Slide 21:
In our first group of experimental tests, we took 10 animals treated with 4 hours of diapulse, started one hour after injury, continuing until the fifth hour after injury when the animals were no longer treated. These are the results 4 months after injury compared to controls.
In terms of ability to walk, none of the control animals, 50% of the pulsed electromagnetic treated animals were able to walk. In terms of vestibular spinal responses, we found 20% of the untreated animals had some vestibular responses, while 60% of the animals had responses in the pulsed field group. SEP's likewise, some of the controls had some responses but the pulsed fields were significantly better. All these three tests are significant to P < .001 using the chi square test. We also wanted to clearly define the axon survival so we did a heroic thing which took nearly a year to do, on the first series of studies. We counted all the axons in the spinal cord.
Slide 22:
The typical rim of the spinal cord after injury, you can see the preservation of myelinated fibers that are most dense close to the surface, becoming very sparse and then you end up with a cavity in the middle of the cord.
Slide 23:
What we did was take the cord and cut it into pie shaped forms. This is a technique of ours devised by Dr. Andrew Blight. What we did was simply draw radial lines for each pie and we camera lucidered through each axon that was encountered along the radial line and put them into a computer. It was heart breaking and back-breaking work. It takes a technician two full weeks to analyze a cord. It took us six months to analyze all the animals in the first series of studies. Let me quickly describe the results.
Slide 24:
What we found in our control animals, virtually all the animals had less than 10,000 fibers remaining in the spinal cord after injury, compared to a total original population of greater than 500,000. 5,000 out of 500,000 is 1% of the total population. Its quite a severe injury we are inflicting. The treatment with Diapulse increased that number, particularly in the animals that walked, in about 30 to 40% of the animals. They overlapped with the control group and we are not sure why the pulsed field therapy did not work in the animals that overlapped.
Here is the data: we have five animals fully counted here, these are the axons numbers in thousands, this is the pulsed electromagnetic field treated group and the solid rectangle or boxes are the animals that are walking. Incidentally, there appears to be a bit of overlap here in this one particular animal. What we found was a very interesting thing. When we limited our count to only the first hundred microns from the surface of the cord and incidentally, 90% of the fibers can be found within the first hundred microns. If I had shown the graft of what it was in the first hundred microns, there is no overlap between the walking versus the non-walking animal. In fact, we are now convinced there is some magic number there, about 25,000 fibers the animal need to walk. Animals that have substantially more than that will walk, animals that have even slightly less than that will not be able to walk.
Slide 25:
Let me quickly run through the experiments we did trying to localize the time course of the effect of the pulsed electromagnetic field.
Slide 26:
The first set of experiments I just described was the control "A" and the PMF treated group "A-1" where the treatment extended from one hour to five hours after injury. The remaining group here were treated, starting at one hour, extending only for one hour. Then starting at four hours going to five hours and all three groups were treated weekly, for two weeks, daily, for only one hour.
Slide 27:
These represent the four month results in this group of animals. In the two control groups, one out of eight "A" animals had SEP's, one out of control "B" had SEP's. Five animals in the "A" group had SEP's. What was very interesting was that we found the best effect to be in the group where we started treatment at four hours after injury extending it for one hour. The results here were very similar to the four hour treatment. Simply by treating this window at four to five hours.
Slide 28:
Here is the data on four months walking in those three groups of animals and again what we found was the best results occurred within this window of treatment. However, we did have some walkers in the 24 four group compared to none in the controls.
Slide 29:
This is the free-fall drop data. Again the best results occurred for this particular treatment paradigm. Unless you think it's such a simple story, I did not present any of the intervening SEP or evoked potential changes. The treatment is actually much more impressive than what it appears at four months. If we looked at the data, for example, for the SEP recovery in these cats, at thirty days after injury, the results are very spectacular. Seventy percent of the PMF treated animals had evoked potentials compared to zero percent of the controls. However, by two months after injury the numbers dropped substantially. The same thing is true for the vestibular spinal response, except for the one hour treated group. We believe that even if the pulsed electromagnetic field is started 24 hours after injury there is a significant effect on the evoked potential and neurophysiological changes in the spinal cord.
Slide 30:
I would like to close here by simply reviewing our finding:
Calcium enters cells after spinal cord injury.
Calcium progressively accumulates in the spinal cord.
Pulsed electromagnetic fields reduce calcium accumulation in the cord.
PMF treatment also improves motorsensory recovery, suggesting that calcium is likely to have a role in spinal cord injury (We were astonished by some of these results).
Finally, I want to point this out; with all the interest in calcium channel blockers and calcium in spinal cord injury as well as other types of injury, we believe pulsed electromagnetic fields may prove to be a useful tool for manipulating CNS calcium.